Since beginning of this week, Amazon has been blocking Meta’s new AI agent Muse from gaining access to its shopping pages. The company had previously tried in vain to persuade Facebook’s parent company to voluntarily refrain from allowing the bot to make purchases on behalf of consumers at its online shop. This highlights just how fiercely companies are fighting for control over their shoppers’ data.
The online giant wants to prevent the AI agent from accessing all information stored in a customer’s account. If the bot logs in as an Amazon customer, for example, it can view a list of past purchases, previous search queries as well as contact and payment details. This is valuable information that offers significant competitive advantages. Market players usually pay high sums for such data.
Amazon is particularly critical of the fact that Meta’s AI assistant does not identify itself as such. This means that a third party could secretly extract sensitive shopper data and carry out transactions without the retailer’s knowledge or consent. Clearly, this is not just a matter of data protection, but above all a question of control over valuable shopper data.
Amazon is also acting against other AI agents
The company explains that other services operating on behalf of customers do identify themselves as such. They act with the store operator’s consent and have confirmed to handle data securely and reliably. Amazon cited delivery services and restaurants, as well as travel agencies and airlines, as examples of this. Agent-based AI applications must adhere to these obligations in the same way.
Muse is not the first agent-based AI that Amazon has acted against. In March this year, the company sued Perplexity under the US Anti-Hacking Act. The online retailer won at first instance but subsequently lost the case on the grounds that it was the user, not the AI company, who was accessing Amazon’s servers. The court did, however, point to the option of controlling such access through contracts or terms of use.
AI threatens Amazon’s dominant position
Amazon also refuses to allow OpenAI’s ChatGPT Shopping Research onto its site. The company has been blocking the tool since November 2025. The primary concern here was not so much fear of data theft. Consumers who use AI agents to search for the best deals may, as a result, become aware of lesser-known and less financially powerful suppliers – including those who cannot buy their way to a top position in Google search results. Amazon sees this as a particular threat, as many consumers have been using its website – due to its market position – as their first port of call when searching for products and, for convenience’s sake, make their purchases there straight away. Yet the online giant is often not the cheapest provider.
However, closing itself off carries the risk that, in future, more consumers will rely on AI tools, meaning fewer customers will end up on Amazon. Should the company become more open, it would lose control of the customer relationship to an external platform. That is why Amazon has become proactive. With ‘Alexa for Shopping’, it has developed its own AI agent for product searches and personalised recommendations. Its ‘Buy for Me’ agent even finds items on third-party product pages. The company assures, however, that the agent identifies itself and offers sellers an opt-out option.
Muse has already reached the top spot
Meta launched Muse on 8 September. Users can utilise the software as a personal assistant. The bot can carry out complex, multi-step tasks and connect to a wide variety of services, such as email, calendars, payment services, restaurant bookings or online shops.
The app is available via its own website as well as on iOS, Android and WhatsApp. The basic version is free of charge. Just one week after its launch, it reached the top spot among free apps on Apple’s US App Store, ahead of competitor ChatGPT.
